Can I use Claude, ChatGPT, or other AI assistants with Loadster?
Yes you can! Loadster has a built-in MCP server, so you can connect your AI assistant to your Loadster account and let it do the slow and boring parts of load testing for you.
Scripting is usually the slowest part of load testing, so an assistant that can read your app, write a script, play it, and fix what breaks tends to speed things up quite a bit. You describe the flow you want to test, and the assistant builds it in your Loadster project using the same tools you’d use in the dashboard.
Which AI Assistants Work with Loadster
Any assistant that supports remote MCP servers over Streamable HTTP and can authenticate with OAuth or a bearer token can connect. Things change quickly, but at the time of writing common options include:
- Claude — both Claude Code and the Claude desktop app are MCP clients.
- ChatGPT — install the official Loadster plugin from the ChatGPT plugin directory.
- Codex — the same Loadster plugin installs in the Codex desktop app and CLI, and the IDE extension connects through Codex’s MCP configuration.
- Cursor and VS Code — access MCP servers from a project or user config file.
Other compatible MCP tools use the same Loadster endpoint, although their setup and authentication options can differ. Loadster’s MCP server exposes its script editor, scenarios, monitors, and results, and the assistant calls those tools on your behalf.
What an AI Assistant Can Do in Loadster
Your assistant can write and play scripts, build load test scenarios, manage datasets and monitors, and read finished results.
What assistants deliberately cannot do is launch a full-scale load test or start a monitor, both of which are left to humans for now to prevent expensive surprises. See also: Can my AI assistant run load tests or start monitors?
Check out the MCP manual page for detailed connection steps for each client.